Re: [PATCH net-next 2/4] net: ethernet: ti: am65-cpsw: streamline .probe() error handling

[Russell King (Oracle)]

On Wed, Jan 15, 2025 at 06:43:01PM +0200, Roger Quadros wrote:
> Keep things simple by explicitly cleaning up on .probe() error
> path or .remove(). Get rid of devm_add/remove_action() usage.
> 
> Rename am65_cpsw_disable_serdes_phy() to
> am65_cpsw_nuss_cleanup_slave_ports() and move it right before
> am65_cpsw_nuss_init_slave_ports().
> 
> Get rid of am65_cpsw_nuss_phylink_cleanup() and introduce
> am65_cpsw_nuss_cleanup_ndevs() right before am65_cpsw_nuss_init_ndevs()
> 
> Move channel initiailzation code out of am65_cpsw_nuss_register_ndevs()
> into new function am65_cpsw_nuss_init_chns().
> Add am65_cpsw_nuss_remove_chns() to do reverse of
> am65_cpsw_nuss_init_chns().
> 
> Add am65_cpsw_nuss_unregister_ndev() to do reverse of
> am65_cpsw_nuss_register_ndevs().
> 
> Use the introduced helpers in probe/remove.

Wow, so we're now saying that devm shouldn't be used? Given that patch 1
is wrong, I'm not sure I'd trust this patch to be correct either as it
goes against what I understand is preferred - to avoid explicit cleanups
that can get in the wrong order or be missed.

-- 
RMK's Patch system: https://www.armlinux.org.uk/developer/patches/
FTTP is here! 80Mbps down 10Mbps up. Decent connectivity at last!